The SM-MIMO communication system is a communication system employing the SM scheme and the MIMO scheme, and various signal detection schemes, such as the ML (Maximum Likelihood) scheme, the ZF (Zero Forcing) scheme, the MMSE (Minimum Mean Square Error) scheme, the OSIC (Order Successive Interference Cancellation) scheme, the sphere decoding scheme, and the QRD-M (QR Decomposition-M), have been suggested in the SM-MIMO communication system.
The ZF scheme and the MMSE scheme have lower complexity, but their performance deterioration is very large in comparison with the ML scheme. Further, the OSIC scheme generates more performance deterioration in comparison with the ML scheme. Further, the sphere decoding scheme has identical performance to the ML scheme, but it has difficulty in obtaining a radius of an initial sphere and it requires very large calculation quantity in consideration of the worst channel condition such that it is difficult to achieve the actual implementation of the sphere decoding scheme. Further, the QRD-M scheme has performance that greatly differs according to the number of candidate groups, such that the multiple candidate groups are required so as to obtain performance identical to that of the ML scheme, and in the event of considering the multiple candidate groups, the calculation complexity in the QRD-M scheme increases.
A signal transmission device of the SM-MIMO communication system uses a channel coding scheme so as to improve the performance and increase the channel capacity. Therefore, a signal reception device corresponding to the signal transmission device generates a LLR (Log-Likelihood Ratio) value for soft decision of a channel decoder.
When the ZF scheme, the MMSE scheme, the OSIC scheme, the sphere decoding scheme, the QRD-M scheme, and the like, are used, the signal reception device calculates a square value of the Euclidean distance of the candidate group for generating the LLR value. Especially, when the signal reception device uses the sphere decoding scheme, the large number of candidate groups of the detected signals are used for the performance improvement and thus the calculation complexity also increases.
In the QRD-M scheme, the Euclidean distances are required in order to generate the LLR value, for both the situation where a value of each bit is 0 and the situation where each bit has a value of 1. However, a corresponding bit may be null in the candidate group, which may make it impossible to detect the LLR value. Various schemes for estimating the Euclidean distance of the null bit have been suggested, but they have still include a limitation in the accuracy aspect, so that it may be impossible to generate the optimum LLR value.
As described above, the signal detection schemes suggested in the SM-MIMO communication system have deteriorated performance and increased calculation complexity in a process of the LLR value generation, as well as a process of the signal detection, in comparison with the ML scheme. Therefore, a signal detecting scheme, which can show a performance similar to that of the ML scheme and decrease the calculation complexity, has been necessarily demanded.